Safety-Aware Image-to-Image Translation without Paired Data
Wonjun Lee
I received my M.S. from Yonsei University, where I worked at CVLab, advised by Prof. Bumsub Ham, and am also affiliated with KDST in Kyung Hee University (formerly at KIST), co-advised by Prof. Suhyun Kim.
I am broadly interested in building trustworthy and safe AI that can operate reliably under real-world conditions.
If you have any questions, feel free to contact me!
โ๏ธ E-mail: velpegor@yonsei.ac.kr
๐ฅ Recent News
- Jan. 2026: One paper accepted at ICLR 2026
- Sep. 2025: New preprint is out: SceneSplit
- May. 2025: One paper accepted at ICML 2025
- Feb. 2025: New preprint is out: ELITE
- Dec. 2024: One paper accepted at AAAI 2025
๐ง Research Interests
๐ Publications
Jailbreaking on Text-to-Video Models via Scene Splitting Strategy
ELITE: Enhanced Language-Image Toxicity Evaluation for Safety
Maximizing the Position Embedding for Vision Transformers with Global Average Pooling
๐ป Experience
Research Collaboration - AIM Intelligence
Student Researcher - Center for Artificial Intelligence, Korea Institute of Science and Technology (KIST)
Research Intern - Center for Artificial Intelligence, Korea Institute of Science and Technology (KIST)
๐ Education
Mar. 2024 - Feb. 2026: M.S. in Electrical and Electronic Engineering, Yonsei University
Advisor: Prof. Suhyun Kim and Prof. Bumsub Ham
Mar. 2018 - Feb. 2024: B.S. in Software Convergence, Kyung Hee University
๐งพ Academic Services
-
Reviewer
The AI for Creative Visual Content Generation, Editing and Understanding (CVEU) Workshop @ CVPR 2026